Skip to content

[fix] handle empty args in chain_matmul #43553

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
kshitij12345 wants to merge 2 commits into from
Closed

[fix] handle empty args in chain_matmul #43553

kshitij12345 wants to merge 2 commits into from

Conversation

@kshitij12345
Copy link
Collaborator

@kshitij12345 kshitij12345 commented Aug 25, 2020

Fixes #41817

@mruberry mruberry added the module: linear algebra Issues related to specialized linear algebra operations in PyTorch; includes matrix multiply matmul label Aug 26, 2020
@mruberry mruberry self-requested a review August 26, 2020 04:24
@mruberry mruberry added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Aug 26, 2020
@mruberry
Copy link
Collaborator

mruberry commented Aug 26, 2020

Hey @kshitij12345! Would you add a test for this behavior to ensure it doesn't regress?

kshitij12345
kshitij12345 commented Aug 26, 2020
View reviewed changes
test/test_torch.py
run_test([10, 20, 30, 5])
run_test([15, 5, 10, 20, 25])

with self.assertRaisesRegex(RuntimeError, "chain_matmul: Expected one or more matrices"):
Copy link
Collaborator Author

@kshitij12345 kshitij12345 Aug 26, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Have updated the existing test here. To verify the behaviour.

Copy link
Collaborator Author

@kshitij12345 kshitij12345 Aug 26, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pytorch/test/test_torch.py

Lines 7224 to 7241 in 5fb5244

@dtypes(torch.double)
def test_chain_matmul(self, device, dtype):
def product(matrices):
for mat in matrices[1:]:
matrices[0] = matrices[0].mm(mat)
return matrices[0]
def run_test(p):
matrices = []
for (pi, pi_1) in zip(p[:-1], p[1:]):
matrices.append(torch.randn(pi, pi_1, dtype=dtype, device=device))
self.assertEqual(torch.chain_matmul(*matrices), product(matrices))
run_test([10, 20, 30, 5])
run_test([15, 5, 10, 20, 25])
with self.assertRaisesRegex(RuntimeError, "chain_matmul: Expected one or more matrices"):
torch.chain_matmul()

mruberry
mruberry approved these changes Aug 26, 2020
View reviewed changes
Copy link
Collaborator

@mruberry mruberry left a comment •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Awesome! Thanks @kshitij12345!

facebook-github-bot
facebook-github-bot reviewed Aug 26, 2020
View reviewed changes
Copy link
Contributor

@facebook-github-bot facebook-github-bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@mruberry has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

@facebook-github-bot
Copy link
Contributor

facebook-github-bot commented Aug 27, 2020

@mruberry merged this pull request in 01b5c06.

@kshitij12345 kshitij12345 deleted the fix/chain-matmul/empty-args branch August 28, 2020 16:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@facebook-github-bot facebook-github-bot facebook-github-bot left review comments

@mruberry mruberry mruberry approved these changes

Assignees

No one assigned

Labels

Merged module: linear algebra Issues related to specialized linear algebra operations in PyTorch; includes matrix multiply matmul open source triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

Segfault when passing an empty input to torch.chain_matmul

4 participants

@kshitij12345 @mruberry @facebook-github-bot @pytorchbot